diff --git a/README.md b/README.md index d560ddaa5115357e08226f200ee6ec9f3ce10af8..bc37152263e5a9e528f3dbf9f6b2b857510e9f43 100644 --- a/README.md +++ b/README.md @@ -31,11 +31,11 @@ datasets/ ### MODELS -Download the trained models at [Onedrive](https://mailstsinghuaeducn-my.sharepoint.com/:f:/g/personal/liuxy21_mails_tsinghua_edu_cn/EgOQzGZn9A5DlaQhgpTtHBwBGWEB57mpYy4SbmpZJMmMyQ?e=0z9Xd2) and put them in the folder ./output. +Download the trained models at [Onedrive](https://mailstsinghuaeducn-my.sharepoint.com/:f:/g/personal/liuxy21_mails_tsinghua_edu_cn/EgOQzGZn9A5DlaQhgpTtHBwB2Bwyx8qmvLauiHFcJbnGSw?e=EZ60La) (password: groupji) and put them in the folder `./output`. ## Requirements -* Ubuntu 16.04/18.04, CUDA 10.1/10.2, python >= 3.6, PyTorch >= 1.6, torchvision +* Ubuntu 16.04/18.04/20.04, CUDA 10.1/10.2, python >= 3.6, PyTorch >= 1.6, torchvision * Install `detectron2` from [source](https://github.com/facebookresearch/detectron2) * `sh scripts/install_deps.sh` * Compile the cpp extension for `farthest points sampling (fps)`: